Vol. 129, No. 2
OF BACTERIOLOGY, Feb. 1977, p. 967-972 Copyright © 1977 American Society for Microbiology
JouRNAL
Printed in U.S.A.
Isolation of an Escherichia coli Mutant Deficient in Thioredoxin Reductase JAMES FUCHS Department of Biochemistry, University of Minnesota, College of Biological Sciences, St. Paul, Minnesota 55108
Received for publication 7 September 1976
A mutant of Escherichia coli defective in thioredoxin reductase has been isolated and partially characterized. This mutant has no detectable thioredoxin reductase activity in vitro and yet it exhibits no in vivo defect in reduction of ribonucleotides. Evidence is presented that indicates that, in cells permeabilized via ether treatment, ribonucleoside diphosphate reduction can utilize glutathione as an alternate reducing system.
Ribonucleoside diphosphate (RDP) reduction in Escherichia coli constitutes the first metabolic reaction unique to deoxyribonucleic acid (DNA) synthesis. This reaction is catalyzed by the enzyme RDP reductase which consists of two nonidentical subunits, proteins B1 and B2, which form an active complex in the presence of magnesium ions (5). The physiological hydrogen donor for this reaction is believed to be reduced nicotinamide adenine dinucleotide phosphate (NADPH). For NADPH to serve as a hydrogen donor, two additional proteins, thioredoxin and thioredoxin reductase, are required (1, 15). Thioredoxin reductase, a flavoprotein that contains an active center oxidation-reduction disulfide, catalyzes the transfer of reducing potential from NADPH to thioredoxin. RDP reductase catalyzes the reduction of ribonucleoside diphosphates using reduced thioredoxin as the hydrogen donor. This reaction was recently shown to involve an oxidation-reduction active disulfide of subunit Bi and to occur by a ping-pong mechanism (22). The overall reaction sequence for the transfer of electrons from NADPH to ribonucleotides is shown in Fig. 1. Brown et al. found that direct reduction of subunit Bi by dithiothreitol can bypass the requirement for this "thioredoxin system"; however, the rate of ribonucleotide reduction was greatly decreased in the absence of thioredoxin (4). The rate of reduction of ribonucleotides using direct reduction of thioredoxin by dithiothreitol was equal to the rate in the presence of thioredoxin, thioredoxin reductase, and NADPH. In conditions more closely approximating physiological conditions, cells made permeable to nucleotides were found to have maximal ribonucleotide reductase activity in the presence ofboth dithiothreitol and NADPH
Thioredoxin systems similar to the E. coli system have been purified from yeast (18), regenerating rat liver cells (11, 12), ascites hepatoma cells (8), and Lactobacillus (17). A different type of thioredoxin system was recently isolated from Euglena (16). In addition to its participation in the reduction of ribonucleotides, the "thioredoxin system" from yeast has been shown to participate in the in vitro reduction of sulfate and in the in vitro reduction of methionine sulfoxide (19). The thioredoxin system isolated from E. coli was able to substitute for the yeast thioredoxin system in these reactions (19). Yeast appear to have separate methionine sulfoxide reductase enzymes for the reduction of the two isomers of L-methionine sulfoxide since partial purification of the enzyme causes a loss of the ability to reduce L-methionine-dsulfoxide (2). Although methionine sulfoxide reductase has not been purified from E. coli, intact cells can reduce all four isomers of methionine sulfoxide (7). Recently, Tsang and Schiff (23) reported that a protein apparently identical to thioredoxin participates in sulfate reduction in E. coli. Host thioredoxin has also been found to constitute one of the subunits of the E. coli bacteriophage T7 DNA polymerase (14). In this paper the isolation of an E. coli mutant deficient in thioredoxin reductase is reported. The mutant does not appear to be defective in ribonucleotide reduction in vivo in spite of having little or no detectable thioredoxin reductase activity in vitro, and utilization of methionine sulfoxide is only slightly diminished. The mutant supports the growth of T7 phage and can use sulfate as its sole sulfur
(24). 967
source. MATERIALS AND METHODS Materials. [3H]uridine 5'-diphosphate was purchased from Schwarz/Mann and was purified before
968
J. BACTERIOL.
FUCHS
10.6 nmol (55 nCi) of L-['4C]methionine-d,l-sulfoxide, and crude E. coli extract. Incubation at 370C for 15 min was followed by heating for 2 min at 100°C to SH H SH stop the reaction. A total of 10 pLI of the reaction mix NA DP B2 NDP was spotted on cellulose-thin layer plates and chroTRose FADH2FIG. 1. Transfer of electrons from NADPH to ri- matographed in a solvent containing isopropanolbonucleotides. TRase, thioredoxin reductase. TR, formic acid-water (80:5.4:20). The areas corresponding to ninhydrin-reacting L-methionine in parallel thioredoxin. Bl *B2, subunits of RDP reductase. channels of the plate were cut out and placed in a scintillation vial. One milliliter of water was added use. The [3H]uridine 5'-diphosphate was concentrated and streaked on a thin-layer plate (20 by 20 to elute the L-[14C]methionine from the cellulose, cm) of polyethyleneimine cellulose and chromato- and 10 ml of a 2:1 mixture of phase-combining sysgraphed as follows: 1 M ammonium acetate contain- tem-xylene scintillation solution was added and ing 2.5% boric acid was allowed to migrate to 1 cm counted in a Beckman LS235 liquid scintillation above the origin, the plate was immediately placed spectrometer. Thioredoxin assay. Thioredoxin assays were conin a solvent of 2 M ammonium acetate containing 5% boric acid, and the solvent was allowed to migrate to ducted in a manner similar to thioredoxin reductase 15 cm above the origin. Solvent salts were removed assays. Thioredoxin and NADPH were omitted from by washing the polyethyleneimine cellulose plate in the.above assay and 32 nmol of dithiothreitol was 100 ml of anhydrous methanol containing 125 mg of added. RDP reductase assay of ether-treated cells. Cells tris(hydroxymethyl)aminomethane and then in 100 ml of anhydrous methanol. The [3H]uridine 5'-di- were grown, harvested, ether treated and assayed as phosphate was eluted with 1 M triethylammonium described previously (6) with the following modificabicarbonate (pH 8.1). The triethylammonium bicar- tions: [3H]uridine 5'-diphosphate was used as substrate at a concentration of 2 mM (1.4 x 104 cpm/ bonate was removed by lyophilizing twice. L-['4C]methionine was purchased from Amer- nmol) and adenosine triphosphate was replaced by sham/Searle Corp. and was converted to L- thymidine triphosphate (0.4 mM). To prevent break['4C]methionine-d,l-sulfoxide by the method of down of the product and to serve as a carrier for Greene (7) except that the procedure was scaled chromatography, 3 mM deoxyuridine 5'-monophosdown 1,000-fold and the precipitates were collected phate was added to the reaction mix. Essentially all by centrifugation rather than by filtering. Greater of the deoxyuridine produced in these assays was than 99.9% of the 14C product co-chromatographed recovered as deoxyuridine 5'-monophosphate presumably due to the action of nucleoside diphosphate with authentic L-methionine-d,l-sulfoxide. A phase-combining system was obtained from kinase and deoxyuridine 5'-triphosphate pyrophosphatase. A total of 10 ,lI of reaction mix was spotted Amersham/Searle. Thioredoxin and thioredoxin reductase were pre- onto polyethyleneimine-impregnated cellulose thinpared from E. coli K-12 as described previously (6). layer plates at 5-min intervals. The deoxyuridine Methionine sulfoxide reductase was prepared 5'-monophosphate was separated from uridine from yeast as described by Block et al. (2) except ribonucleotides by chromatography in a solvent conthat the product of their purification scheme was taining 30 ml of ethylene glycol, 70 ml of water, 6 g further purified by gel filtration on Sephadex G-50. of sodium borate and 3 g of boric acid. The deoxyuriThe methionine sulfoxide reductase was free of any dine 5'-monophosphate spot was cut out, eluted, and detectable yeast thioredoxin or thioredoxin reduc- counted as described previously (6). DNA synthesis. An exponential culture was tase activity. Bacterial strains. The strains used in this study grown in minimal media containing methionine (30 ,g/ml), deoxyadenosine (100 ,ug/ml), and are derivatives ofE. coli KK1004 (6) which is metB [14C]thymidine (2 ug/ml [0.5 uCi/ml]). At indicated upp udk thyA (Ts). KK1006 used as the parental strain in this study was obtained from strain times, 50 ,ul of culture was transferred to 1-inch KK1004 without mutagenesis by selecting a deriva- (2.54-cm) diameter disks of Whatman no. 1 filter tive that would form distinct colonies on the back- paper, and the disks were immersed in 5% trichloroground growth that occurs in minimal agar supple- acetic acid containing 1% Na4P207 10H20. The disks mented with 2 ,ug of thymine per ml. Strain KK1006 were washed twice in the trichloroacetic acid, then is not sensitive to deoxyadenosine and thus is pre- twice in 95% ethanol, and then twice in ether, and air dried (3). The disks were counted in a scintillasumably a drm derivative of strain KK1004. Growth of bacteria. The media used and growth tion solution containing 0.4% 2,5-diphenyloxazole of cells were as described previously (6). When and 0.01% 1,4-bis[2]-(5-phenyloxazolyl)benzene in methionine sulfoxide was used in place of methi- toluene. onine, it was at a concentration of 30 ,ug/ml. Mutagenesis and penicillin treatment were carried out as RESULTS described by Karlstrom (10). Thioredoxin reductase assays. Thioredoxin reRationale for selection. Exogenous deoxyductase activity was determined in a coupled assay. In a total volume of 25 p.l the following components nucleosides cannot be utilized by E. coli for DNA synthesis, with the exception of thymidine and were added: 2.4 ug of yeast L-methionine-l-sulfoxide reductase, 250 pmol of E. coli thioredoxin, 2.5 u.l of deoxyuridine, due to the absence of deoxynu0.5 M Tris-chloride (pH 7.7), 10 nmol of NADPH, cleoside kinases. Thus, mutants unable to synS-S
NADPNHHy
TRose FAD
NADP0SHOTRosexFAD
S-S
VTR''
SN
,TR
\ /
B1B2dNDP
VOL. 129, 1977
THIOREDOXIN REDUCTASE-DEFICIENT E. COLI 969 thesize deoxynucleotides can only be obtained as conditional mutants. Since E. coli appeared to require thioredoxin and thioredoxin reductase for ribonucleotide reduction as well as for KKK 1006 E the reduction of methionine sulfoxide, condiw z tional mutants deficient in thioredoxin or thioredoxin reductase were sought. In a methionine 0 auxotrophic strain, a mutant with a partially defective thioredoxin or thioredoxin reductase, IV that had sufficient activity for ribonucleotide reduction but not sufficient activity to simultaE neously support both reduction of ribonucleotides and the methionine sulfoxide required for 105 40 501 growth, could exist. Such a mutant might be expected to grow when supplied with methioKK 10448 nine rather than with methionine sulfoxide, 0 --? thus sparing the thioredoxin system for use in O ISO 50 100 ribonucleotide reduction. If such a mutant had ^g CRUDE EXTRACT PROTEIN a defective thioredoxin reductase with inFIG. 2. Thioredoxin reductase activity in crude creased thermolability, no growth would be ex- extracts of strains KK1006 and KK1048. Incubation pected at higher temperatures even on enriched mixture contained yeast methionine sulfoxide reducmedia. tase, E. coli thioredoxin, ['4Cimethionine sulfoxide, The methionine auxotroph, KK1006, was and NADPH as reducing agent. Symbols: (@) strain mutagenized with N-methyl-N'-nitro-N-nitroso- KK1006; (0) strain KKZ048. guanidine, phenotypically expressed and then subjected to counterselection by penicillin in reductase activity at various minimal media containing methionine sulfox- TABLE 1. Thioredoxin temperatures ide as the sole source of methionine. Three hundred survivors were tested for their ability Strain Temp Activity" _C) to grow on minimal media containing either 25 2.48 methionine or methionine sulfoxide. Three mu- KK1006 25
OF BACTERIOLOGY, Feb. 1977, p. 967-972 Copyright © 1977 American Society for Microbiology
JouRNAL
Printed in U.S.A.
Isolation of an Escherichia coli Mutant Deficient in Thioredoxin Reductase JAMES FUCHS Department of Biochemistry, University of Minnesota, College of Biological Sciences, St. Paul, Minnesota 55108
Received for publication 7 September 1976
A mutant of Escherichia coli defective in thioredoxin reductase has been isolated and partially characterized. This mutant has no detectable thioredoxin reductase activity in vitro and yet it exhibits no in vivo defect in reduction of ribonucleotides. Evidence is presented that indicates that, in cells permeabilized via ether treatment, ribonucleoside diphosphate reduction can utilize glutathione as an alternate reducing system.
Ribonucleoside diphosphate (RDP) reduction in Escherichia coli constitutes the first metabolic reaction unique to deoxyribonucleic acid (DNA) synthesis. This reaction is catalyzed by the enzyme RDP reductase which consists of two nonidentical subunits, proteins B1 and B2, which form an active complex in the presence of magnesium ions (5). The physiological hydrogen donor for this reaction is believed to be reduced nicotinamide adenine dinucleotide phosphate (NADPH). For NADPH to serve as a hydrogen donor, two additional proteins, thioredoxin and thioredoxin reductase, are required (1, 15). Thioredoxin reductase, a flavoprotein that contains an active center oxidation-reduction disulfide, catalyzes the transfer of reducing potential from NADPH to thioredoxin. RDP reductase catalyzes the reduction of ribonucleoside diphosphates using reduced thioredoxin as the hydrogen donor. This reaction was recently shown to involve an oxidation-reduction active disulfide of subunit Bi and to occur by a ping-pong mechanism (22). The overall reaction sequence for the transfer of electrons from NADPH to ribonucleotides is shown in Fig. 1. Brown et al. found that direct reduction of subunit Bi by dithiothreitol can bypass the requirement for this "thioredoxin system"; however, the rate of ribonucleotide reduction was greatly decreased in the absence of thioredoxin (4). The rate of reduction of ribonucleotides using direct reduction of thioredoxin by dithiothreitol was equal to the rate in the presence of thioredoxin, thioredoxin reductase, and NADPH. In conditions more closely approximating physiological conditions, cells made permeable to nucleotides were found to have maximal ribonucleotide reductase activity in the presence ofboth dithiothreitol and NADPH
Thioredoxin systems similar to the E. coli system have been purified from yeast (18), regenerating rat liver cells (11, 12), ascites hepatoma cells (8), and Lactobacillus (17). A different type of thioredoxin system was recently isolated from Euglena (16). In addition to its participation in the reduction of ribonucleotides, the "thioredoxin system" from yeast has been shown to participate in the in vitro reduction of sulfate and in the in vitro reduction of methionine sulfoxide (19). The thioredoxin system isolated from E. coli was able to substitute for the yeast thioredoxin system in these reactions (19). Yeast appear to have separate methionine sulfoxide reductase enzymes for the reduction of the two isomers of L-methionine sulfoxide since partial purification of the enzyme causes a loss of the ability to reduce L-methionine-dsulfoxide (2). Although methionine sulfoxide reductase has not been purified from E. coli, intact cells can reduce all four isomers of methionine sulfoxide (7). Recently, Tsang and Schiff (23) reported that a protein apparently identical to thioredoxin participates in sulfate reduction in E. coli. Host thioredoxin has also been found to constitute one of the subunits of the E. coli bacteriophage T7 DNA polymerase (14). In this paper the isolation of an E. coli mutant deficient in thioredoxin reductase is reported. The mutant does not appear to be defective in ribonucleotide reduction in vivo in spite of having little or no detectable thioredoxin reductase activity in vitro, and utilization of methionine sulfoxide is only slightly diminished. The mutant supports the growth of T7 phage and can use sulfate as its sole sulfur
(24). 967
source. MATERIALS AND METHODS Materials. [3H]uridine 5'-diphosphate was purchased from Schwarz/Mann and was purified before
968
J. BACTERIOL.
FUCHS
10.6 nmol (55 nCi) of L-['4C]methionine-d,l-sulfoxide, and crude E. coli extract. Incubation at 370C for 15 min was followed by heating for 2 min at 100°C to SH H SH stop the reaction. A total of 10 pLI of the reaction mix NA DP B2 NDP was spotted on cellulose-thin layer plates and chroTRose FADH2FIG. 1. Transfer of electrons from NADPH to ri- matographed in a solvent containing isopropanolbonucleotides. TRase, thioredoxin reductase. TR, formic acid-water (80:5.4:20). The areas corresponding to ninhydrin-reacting L-methionine in parallel thioredoxin. Bl *B2, subunits of RDP reductase. channels of the plate were cut out and placed in a scintillation vial. One milliliter of water was added use. The [3H]uridine 5'-diphosphate was concentrated and streaked on a thin-layer plate (20 by 20 to elute the L-[14C]methionine from the cellulose, cm) of polyethyleneimine cellulose and chromato- and 10 ml of a 2:1 mixture of phase-combining sysgraphed as follows: 1 M ammonium acetate contain- tem-xylene scintillation solution was added and ing 2.5% boric acid was allowed to migrate to 1 cm counted in a Beckman LS235 liquid scintillation above the origin, the plate was immediately placed spectrometer. Thioredoxin assay. Thioredoxin assays were conin a solvent of 2 M ammonium acetate containing 5% boric acid, and the solvent was allowed to migrate to ducted in a manner similar to thioredoxin reductase 15 cm above the origin. Solvent salts were removed assays. Thioredoxin and NADPH were omitted from by washing the polyethyleneimine cellulose plate in the.above assay and 32 nmol of dithiothreitol was 100 ml of anhydrous methanol containing 125 mg of added. RDP reductase assay of ether-treated cells. Cells tris(hydroxymethyl)aminomethane and then in 100 ml of anhydrous methanol. The [3H]uridine 5'-di- were grown, harvested, ether treated and assayed as phosphate was eluted with 1 M triethylammonium described previously (6) with the following modificabicarbonate (pH 8.1). The triethylammonium bicar- tions: [3H]uridine 5'-diphosphate was used as substrate at a concentration of 2 mM (1.4 x 104 cpm/ bonate was removed by lyophilizing twice. L-['4C]methionine was purchased from Amer- nmol) and adenosine triphosphate was replaced by sham/Searle Corp. and was converted to L- thymidine triphosphate (0.4 mM). To prevent break['4C]methionine-d,l-sulfoxide by the method of down of the product and to serve as a carrier for Greene (7) except that the procedure was scaled chromatography, 3 mM deoxyuridine 5'-monophosdown 1,000-fold and the precipitates were collected phate was added to the reaction mix. Essentially all by centrifugation rather than by filtering. Greater of the deoxyuridine produced in these assays was than 99.9% of the 14C product co-chromatographed recovered as deoxyuridine 5'-monophosphate presumably due to the action of nucleoside diphosphate with authentic L-methionine-d,l-sulfoxide. A phase-combining system was obtained from kinase and deoxyuridine 5'-triphosphate pyrophosphatase. A total of 10 ,lI of reaction mix was spotted Amersham/Searle. Thioredoxin and thioredoxin reductase were pre- onto polyethyleneimine-impregnated cellulose thinpared from E. coli K-12 as described previously (6). layer plates at 5-min intervals. The deoxyuridine Methionine sulfoxide reductase was prepared 5'-monophosphate was separated from uridine from yeast as described by Block et al. (2) except ribonucleotides by chromatography in a solvent conthat the product of their purification scheme was taining 30 ml of ethylene glycol, 70 ml of water, 6 g further purified by gel filtration on Sephadex G-50. of sodium borate and 3 g of boric acid. The deoxyuriThe methionine sulfoxide reductase was free of any dine 5'-monophosphate spot was cut out, eluted, and detectable yeast thioredoxin or thioredoxin reduc- counted as described previously (6). DNA synthesis. An exponential culture was tase activity. Bacterial strains. The strains used in this study grown in minimal media containing methionine (30 ,g/ml), deoxyadenosine (100 ,ug/ml), and are derivatives ofE. coli KK1004 (6) which is metB [14C]thymidine (2 ug/ml [0.5 uCi/ml]). At indicated upp udk thyA (Ts). KK1006 used as the parental strain in this study was obtained from strain times, 50 ,ul of culture was transferred to 1-inch KK1004 without mutagenesis by selecting a deriva- (2.54-cm) diameter disks of Whatman no. 1 filter tive that would form distinct colonies on the back- paper, and the disks were immersed in 5% trichloroground growth that occurs in minimal agar supple- acetic acid containing 1% Na4P207 10H20. The disks mented with 2 ,ug of thymine per ml. Strain KK1006 were washed twice in the trichloroacetic acid, then is not sensitive to deoxyadenosine and thus is pre- twice in 95% ethanol, and then twice in ether, and air dried (3). The disks were counted in a scintillasumably a drm derivative of strain KK1004. Growth of bacteria. The media used and growth tion solution containing 0.4% 2,5-diphenyloxazole of cells were as described previously (6). When and 0.01% 1,4-bis[2]-(5-phenyloxazolyl)benzene in methionine sulfoxide was used in place of methi- toluene. onine, it was at a concentration of 30 ,ug/ml. Mutagenesis and penicillin treatment were carried out as RESULTS described by Karlstrom (10). Thioredoxin reductase assays. Thioredoxin reRationale for selection. Exogenous deoxyductase activity was determined in a coupled assay. In a total volume of 25 p.l the following components nucleosides cannot be utilized by E. coli for DNA synthesis, with the exception of thymidine and were added: 2.4 ug of yeast L-methionine-l-sulfoxide reductase, 250 pmol of E. coli thioredoxin, 2.5 u.l of deoxyuridine, due to the absence of deoxynu0.5 M Tris-chloride (pH 7.7), 10 nmol of NADPH, cleoside kinases. Thus, mutants unable to synS-S
NADPNHHy
TRose FAD
NADP0SHOTRosexFAD
S-S
VTR''
SN
,TR
\ /
B1B2dNDP
VOL. 129, 1977
THIOREDOXIN REDUCTASE-DEFICIENT E. COLI 969 thesize deoxynucleotides can only be obtained as conditional mutants. Since E. coli appeared to require thioredoxin and thioredoxin reductase for ribonucleotide reduction as well as for KKK 1006 E the reduction of methionine sulfoxide, condiw z tional mutants deficient in thioredoxin or thioredoxin reductase were sought. In a methionine 0 auxotrophic strain, a mutant with a partially defective thioredoxin or thioredoxin reductase, IV that had sufficient activity for ribonucleotide reduction but not sufficient activity to simultaE neously support both reduction of ribonucleotides and the methionine sulfoxide required for 105 40 501 growth, could exist. Such a mutant might be expected to grow when supplied with methioKK 10448 nine rather than with methionine sulfoxide, 0 --? thus sparing the thioredoxin system for use in O ISO 50 100 ribonucleotide reduction. If such a mutant had ^g CRUDE EXTRACT PROTEIN a defective thioredoxin reductase with inFIG. 2. Thioredoxin reductase activity in crude creased thermolability, no growth would be ex- extracts of strains KK1006 and KK1048. Incubation pected at higher temperatures even on enriched mixture contained yeast methionine sulfoxide reducmedia. tase, E. coli thioredoxin, ['4Cimethionine sulfoxide, The methionine auxotroph, KK1006, was and NADPH as reducing agent. Symbols: (@) strain mutagenized with N-methyl-N'-nitro-N-nitroso- KK1006; (0) strain KKZ048. guanidine, phenotypically expressed and then subjected to counterselection by penicillin in reductase activity at various minimal media containing methionine sulfox- TABLE 1. Thioredoxin temperatures ide as the sole source of methionine. Three hundred survivors were tested for their ability Strain Temp Activity" _C) to grow on minimal media containing either 25 2.48 methionine or methionine sulfoxide. Three mu- KK1006 25
